Why is it a poor practice for analysts to manage access by copying data into CSV extracts they share among themselves?
Select an answer to reveal the explanation.
Short Explanation
The moment data becomes a shared CSV extract, it leaves the governed layer, so policy, lineage, and audit stop applying. You can't revoke it, you can't track who has it, and it goes stale as soon as the source changes.
Full Explanation
The correct answer is the option stating that copies leave the governed layer, so policy, lineage and audit no longer apply, and they go stale relative to the source. A copied extract is outside the governance boundary. Access policy cannot be enforced on it, revocation is impossible because you do not know where the copies are, lineage no longer describes the data in use, audit no longer applies to the copies, and the extract diverges from the source as soon as the source changes. The option claiming CSV files cannot hold enough rows for analytical work is incorrect because row capacity is not the objection, and it would not matter if the format scaled well. The option claiming the platform blocks any export of table data to a file is incorrect because exports are a normal, supported capability; the problem is governance of what happens afterwards, not the mechanism. The option claiming CSV extracts are slower to read than tables, so queries take longer is incorrect because performance is beside the point; a fast ungoverned copy is still an ungoverned copy. For the exam, the objection to shared extracts is always governance: no policy, no revocation, no lineage, and drift from the source.